AI enablement for investment firms

Your firm's standards, built into the AI your team already uses.

Your templates, your past work, your standards. We build them into the AI tools your firm already pays for, so the whole team produces work your way, with sources to check and a person signing off.

Book a 30-minute call See the case study

For: Private credit managers, private equity firms, family offices, and specialty lenders
Works in: Claude or ChatGPT, under your firm's own account and controls
Recent work: A deal screening and diligence workflow for a private credit manager →

Case study · Private credit

A deal screening and diligence workflow for a private credit manager

A U.S. private credit manager. Every new deal needed a screening summary in the firm's own Excel template and a diligence request list. Both were built by hand, and each person used AI differently.

Scope: Screening summary and diligence request list, built into the firm's own AI account
Approval: Access terms reviewed and approved by the firm's compliance team and counsel
Training: Five live, recorded sessions for the investment team
Handoff: The workflow, a playbook, and the training recordings, all owned by the firm

What happened

The team was testing the first working version within three days and sent structured feedback.
Their feedback shaped three more versions, including updates to the firm's template and analysis the associates had started building themselves.
Associates began proposing their own additions, which we folded into the shared workflow.
Other teams asked for the materials.

deal-screen.xlsx · Sample deal (fictional)Draft · for analyst review

FieldDraftSource

Business overviewCommercial HVAC services, 14 branchesCIM p. 4

Adj. EBITDA$18.2M (LTM)QoE, tab 2

Customer concentrationTop customer is 11% of revenueCustomer file

Key risksTechnician retention, end-market cyclicalityCIM p. 18

Illustration with a fictional company. Drafts land in the firm's own template, with a source for each field, for an analyst to review.

The gap

Everyone has AI. Few firms get consistent work out of it.

Out of the box, these tools don't know your templates, your past work, or how your investment committee likes things presented.

So each person uses them differently, the best methods stay with one person, and the output is hard to check. Individuals get a little faster. The firm doesn't get better.

Off-the-shelf AI	Set up for your firm
Starts from a blank prompt	Starts from your templates and past work
Answers in a chat window	Drafts in your templates
Hard to tell where a number came from	Extracted figures cite the source document and page
Results depend on who is using it	One shared process, with a review checklist

Where it helps

Built around work your firm already does.

Repeatable work with a standard format and a review step is where this pays off first.

Screening and IC memos

First drafts in your template, built from the CIM, financials, and data room.

Diligence request lists

Standard and deal-specific questions for sponsors and management teams, modeled on lists you've sent before.

Financial spreading

Figures from CIMs, QoEs, and borrower financials pulled into your spreadsheets, each one sourced.

Portfolio monitoring

Borrower reporting packages and compliance certificates summarized against budget, prior periods, and covenant levels.

Investor communications

Investor letters, quarterly updates, and DDQ responses, drafted in your firm's voice.

How we work together

Start where your team is today.

Each stage is a fixed fee, scoped in writing before we start. You decide on the next stage after seeing the last one.

Still deciding how to use AI

Workshop

One live session, about an hour

For your leadership or investment team
A live demo on a sample deal
A shortlist of workflows worth building
Help choosing and setting up the right AI plan

Book a workshop

Ready to make one workflow standard

Pilot

About five weeks

One workflow built on your templates and past work
Working versions your team tests on real deals
Five live training sessions, recorded
Playbook, review checklist, and handoff

Talk about a pilot

Ready for the next one

Ongoing

Monthly

Workflows kept current as tools and needs change
Your team's own ideas built in
The next workflow, scoped and built

Book a call about this

What we need from your team

One person who owns the workflow and makes the calls
One or two people who do the work today, to walk us through it and test
Someone to handle access and IT
A handful of completed examples with their source materials

AI drafts. Your team decides.

We'll also tell you where AI isn't reliable enough yet, and keep that work out of scope until it is.

Security and compliance

The answers your compliance team will ask for.

NDAs are standard, and we'll complete your vendor due diligence questionnaire.

Who can see what

We only use the accounts, files, and tools you approve and set up for us. That boundary is written into the statement of work.

Where the work runs

The finished workflow runs in your firm's own AI account, under the data and retention controls you already have.

Model training

We set you up on business plans whose terms keep your data out of model training by default.

Files you share with us

Used only for this engagement and deleted at handoff. We confirm it in writing.

When we're done

Our access is removed at handoff, and we confirm it in writing.

Sensitive information

We follow your policies on material non-public information. Every output is a draft that a person reviews before it's used.

In our most recent engagement, the client's compliance team and counsel reviewed and approved the access terms before work began.

Why Giantfish

Built with the discipline of systems where errors cost money.

Our background is production systems at Mastercard, Walmart, and Disney+, where an error costs money and gets noticed. We bring the same discipline here: sources you can check, code where figures must be exact, and the same senior people from the first call to the handoff.

Questions

How much of our team's time does this take?

One person who owns the workflow, one or two people who do the work today, and someone for access. They walk us through the current process, test working versions on real deals, and join five one-hour training sessions.

How is this priced?

Each stage is a fixed fee, scoped in writing before we start. We'll give you a number after the first call, once we understand the workflow.

We use Microsoft Copilot. Can you work with that?

Often, yes. The approach is the same: your templates, examples, and review steps. What can be automated depends on your licenses and setup, and we'll tell you honestly on the first call.

Do we need Claude or ChatGPT already?

No. If you're still choosing, we'll help you compare plans and set one up that your compliance team is comfortable with.

Who from Giantfish does the work?

Senior people, from the first call to the handoff. We don't pass your work to a junior team.

What happens to our data when you're done?

The workflow stays in your account and belongs to you. We remove our access, delete working files, and confirm both in writing.

Who maintains it if our internal champion leaves?

You get a written playbook and the training recordings, so the next person can pick it up. Ongoing support is there if you'd rather we keep it current.

Are we locked into one AI vendor?

No. Your templates, examples, and review steps carry over between Claude and ChatGPT, and we can help with a switch.

Is this a library of prompts?

No. It's a working process with your templates, built-in checks, and a playbook.

What happens after the pilot?

You own the workflow, the playbook, and the training recordings. Some teams run it on their own from there. Others keep us on for the next workflow.

Start with one workflow.

A 30-minute call. Tell us how your team works today, and we'll tell you honestly where AI can help and where it can't.

Book a 30-minute call Or send a message

Get in touch

Tell us about your firm. We reply within one business day.

Location: New York, NY
Telephone: (646) 470-3551
Email: contact@giantfish.io